Embodiments of the present disclosure relate to generating velocity profiles for an autonomous vehicle (101). An ECU (107) of the autonomous vehicle (101) receives road information from one or more sensors (106) associated with the autonomous vehicle (101). One or more parameters related to smooth movement of the autonomous vehicle on the road is determined from the road information. Further, a first velocity profile is produced using an AI model and a second velocity profile is produced using a hierarchical model, based on the one or more parameters. Furthermore, one of the first and the second velocity profile is selected by comparing the first and the second velocity profiles. The selected velocity profile has a lower value of velocity value compared to the other velocity profile. The selected velocity profile is provided to the autonomous vehicle (101) for navigating on the road (102) smoothly.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
2. The method of claim 1, wherein the road information comprises at least one of, a speed breaker, an inclination, a turn, a pothole, a road boundary, an obstacle, a traffic sign, a road sign, and a median.
This invention relates to systems for detecting and processing road information to enhance vehicle navigation and safety. The technology addresses the problem of accurately identifying and categorizing various road features that impact driving conditions, such as speed bumps, inclines, turns, potholes, road boundaries, obstacles, traffic signs, road signs, and medians. These features are detected using sensors, such as cameras, LiDAR, or radar, and processed to generate real-time data for navigation systems, autonomous driving, or driver assistance systems. The detected road information is used to adjust vehicle speed, trajectory, or alerts to improve safety and efficiency. The system may also integrate with mapping databases to update road conditions dynamically. By identifying and classifying these road features, the invention aims to provide more precise navigation guidance, reduce accidents, and enhance overall driving experience. The technology is particularly useful for autonomous vehicles, advanced driver-assistance systems (ADAS), and smart infrastructure applications.
3. The method of claim 1, wherein the AI model is trained to indicate the most relevant velocity profile from a plurality of velocity profiles using the one or more road parameters provided as input.
This invention relates to an artificial intelligence (AI) system for optimizing vehicle velocity profiles based on road parameters. The problem addressed is the need for adaptive vehicle control that adjusts speed and acceleration patterns in real-time to improve efficiency, safety, or comfort, depending on varying road conditions. The system uses an AI model trained to analyze road parameters such as curvature, slope, surface conditions, traffic density, and weather. These inputs are processed to determine the most relevant velocity profile from a predefined set of profiles. The selected profile dictates the vehicle's speed and acceleration over time, ensuring optimal performance under the given conditions. For example, on a steep downhill road, the model may prioritize a profile that balances speed with braking efficiency to prevent overheating. In urban traffic, it might favor a profile that minimizes stops and starts to reduce fuel consumption. The AI model is trained using historical data and real-time inputs to refine its selection process. The system continuously updates the velocity profile as road parameters change, ensuring dynamic adaptation. This approach enhances vehicle control by replacing static, rule-based systems with a data-driven, context-aware solution. The invention is applicable to autonomous vehicles, advanced driver-assistance systems (ADAS), and fleet management for logistics optimization.
5. The method of claim 1, wherein comparing comprises selecting a lesser velocity value among the first velocity profile and the second velocity profile when a difference is above a threshold value.
This invention relates to a method for comparing velocity profiles in a system where multiple velocity profiles are generated, such as in motion control or robotics applications. The problem addressed is ensuring accurate and reliable velocity comparisons, particularly when discrepancies between profiles may lead to errors in system performance. The method involves generating a first velocity profile based on a first set of parameters and a second velocity profile based on a second set of parameters. These profiles represent the expected or measured velocity of a moving component over time. The method then compares the two profiles to determine their similarity or divergence. If the difference between the profiles exceeds a predefined threshold value, the method selects the lesser velocity value from the two profiles at each comparison point. This ensures that the system operates within safe or optimal limits, preventing excessive speed or instability. The selection of the lesser velocity value helps mitigate risks associated with high-speed discrepancies, such as mechanical stress, control errors, or safety hazards. The threshold value can be dynamically adjusted based on system requirements or environmental conditions. This approach enhances the robustness of velocity-based control systems by ensuring conservative yet reliable operation. The method is particularly useful in applications where precise motion control is critical, such as industrial automation, robotics, or autonomous vehicle navigation.
6. The method of claim 5, wherein the one or more road parameters and the selected velocity profile from the first velocity profile are fed back to enhance the AI model upon the difference between the first velocity profile and the second velocity profile is above the threshold value, wherein the AI model updates one or more model parameters based on the feedback.
This invention relates to autonomous vehicle control systems, specifically improving AI-based velocity profile generation for navigation. The problem addressed is ensuring accurate and adaptive velocity profiles for vehicle movement by refining AI models through feedback mechanisms. The system involves generating a first velocity profile for a vehicle using an AI model based on road parameters such as curvature, slope, and traffic conditions. A second velocity profile is then derived from sensor data or other sources. If the difference between the first and second profiles exceeds a predefined threshold, the road parameters and the selected velocity profile are fed back to the AI model. The AI model then updates its internal parameters to improve future velocity profile predictions. This feedback loop enhances the model's accuracy over time by continuously refining its understanding of road conditions and optimal vehicle behavior. The invention ensures that the AI model adapts dynamically to real-world driving scenarios, reducing discrepancies between predicted and actual velocity profiles. This improves autonomous vehicle safety and efficiency by ensuring the AI model remains aligned with real-time conditions. The feedback mechanism allows the system to learn from deviations, making it more reliable for long-term operation.
7. The non-transitory computer readable medium of claim 1, wherein the AI model is trained to indicate the most relevant velocity profile from a plurality of velocity profiles using the one or more road parameters provided as input.
This invention relates to artificial intelligence models and their application in vehicle systems. Specifically, it addresses the problem of efficiently and accurately selecting the most appropriate velocity profile for a vehicle based on real-world driving conditions. The disclosed technology involves a non-transitory computer-readable medium storing an artificial intelligence (AI) model. This AI model is designed to receive one or more road parameters as input. These road parameters can include information about the road surface, inclines, curves, or other relevant environmental factors. The AI model is trained to analyze these input road parameters and, based on its training, determine and indicate the most relevant velocity profile from a pre-defined plurality of velocity profiles. This allows for dynamic and context-aware velocity selection to optimize vehicle performance, efficiency, or safety.
9. The ECU of claim 8, wherein the processor trains the AI model to indicate the most relevant velocity profile from a plurality of velocity profiles using the one or more road parameters provided as input.
This invention relates to an electronic control unit (ECU) for optimizing vehicle velocity profiles based on road conditions. The system addresses the challenge of improving vehicle efficiency, safety, and comfort by dynamically selecting the most appropriate velocity profile for a given road segment. The ECU includes a processor that trains an artificial intelligence (AI) model to analyze road parameters such as curvature, slope, traffic conditions, and surface quality. The trained model evaluates these inputs to determine the optimal velocity profile from a predefined set of profiles, ensuring smooth acceleration, braking, and speed adjustments tailored to the road environment. The AI model is continuously updated with new data to refine its predictions, enhancing real-time decision-making. This approach reduces fuel consumption, minimizes driver fatigue, and improves overall driving dynamics by adapting to varying road conditions. The system integrates with vehicle sensors and navigation data to provide real-time feedback, ensuring seamless integration with existing automotive control systems. The invention aims to enhance autonomous and semi-autonomous driving capabilities by leveraging AI-driven velocity optimization.
11. The ECU of claim 8, wherein the processor compares by selecting a lesser velocity value among the first velocity profile and the second velocity profile when a difference is above a threshold value.
This invention relates to an electronic control unit (ECU) for managing vehicle dynamics, particularly focusing on velocity profile comparison to enhance control accuracy. The problem addressed is ensuring reliable velocity measurements in dynamic environments where sensor data may be inconsistent or noisy, leading to inaccurate control decisions. The ECU includes a processor that compares two velocity profiles—one derived from a primary sensor and another from a secondary sensor or estimation model. When the difference between these profiles exceeds a predefined threshold, the processor selects the lower velocity value to prioritize safety and conservative control. This approach mitigates potential errors from sensor inaccuracies or transient disturbances, ensuring stable vehicle operation. The ECU may also include additional features such as filtering mechanisms to refine velocity data before comparison, adaptive threshold adjustment based on operating conditions, and integration with other vehicle systems for comprehensive control. The selection of the lesser velocity value helps prevent overestimation of speed, reducing risks like excessive braking or unstable maneuvers. This solution is particularly useful in autonomous driving, advanced driver-assistance systems (ADAS), and other applications requiring precise velocity management under varying conditions. The invention improves reliability by dynamically selecting the most conservative velocity input, enhancing overall system safety and performance.
12. The ECU of claim 11, wherein the processor feedback the one or more road parameters and the selected velocity profile from the first velocity profile to enhance the AI model upon the difference between the first velocity profile and the second velocity profile is above the threshold value, wherein the AI model updates one or more model parameters based on the feedback.
Automotive control systems and artificial intelligence (AI) model enhancement. This invention addresses the problem of improving the accuracy and adaptability of AI models used in vehicle control by incorporating real-time road conditions and driver behavior. An Electronic Control Unit (ECU) is described, which includes a processor. This processor is configured to receive information related to one or more road parameters. It also receives a selected velocity profile, derived from an initial velocity profile. The processor then compares this selected velocity profile with a second velocity profile, likely representing actual vehicle behavior or a target. If the difference between these two velocity profiles exceeds a predefined threshold value, the processor feeds back both the road parameters and the selected velocity profile to enhance the AI model. This feedback mechanism allows the AI model to update its internal parameters, thereby improving its ability to predict and control vehicle speed based on the prevailing road conditions and desired driving characteristics.
14. The non-transitory computer readable medium of claim 13, wherein the road information comprises at least one of, a speed breaker, an inclination, a turn, a pothole, a road boundary, an obstacle, a traffic sign, a road sign, and a median.
This invention relates to a system for enhancing vehicle navigation and safety by processing road information to improve driving assistance. The system captures and analyzes road conditions in real-time or from stored data to identify and classify various road features that impact driving. These features include speed breakers, inclines, turns, potholes, road boundaries, obstacles, traffic signs, road signs, and medians. The system uses this information to generate alerts, adjust navigation routes, or provide driver assistance to avoid hazards or optimize driving behavior. The road information is processed using computer vision, sensor data, or crowdsourced inputs, and the system may integrate with vehicle control systems to automate responses. The goal is to improve safety, efficiency, and comfort by dynamically adapting to road conditions. The invention may be implemented in vehicles, mobile devices, or cloud-based platforms to provide real-time or predictive road condition analysis.
16. The non-transitory computer readable medium of claim 13, wherein comparing comprises selecting a lesser velocity value among the first velocity profile and the second velocity profile when a difference is above a threshold value, wherein the one or more road parameters and the selected velocity profile from the first velocity profile are fed back to enhance the AI model upon the difference between the first velocity profile and the second velocity profile is above the threshold value, wherein the AI model updates one or more model parameters based on the feedback.
This technology relates to vehicular systems and addresses the problem of refining predictive speed profiles for autonomous driving. The invention is a non-transitory computer-readable medium storing instructions that, when executed, enable a system to enhance an Artificial Intelligence (AI) model. The process involves comparing a first velocity profile with a second velocity profile. If the difference between these profiles exceeds a predefined threshold, a lesser velocity value is selected from the two profiles. Crucially, one or more road parameters, along with the selected velocity profile from the first velocity profile, are then utilized as feedback to improve the AI model. This feedback mechanism allows the AI model to update its internal parameters, thereby refining its ability to predict and generate optimal velocity profiles for varying road conditions. This iterative enhancement process leads to more accurate and responsive autonomous driving behavior.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
March 30, 2020
December 6, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.